Strange resets on new computer

ricno

Distinguished
Apr 5, 2010
582
0
19,010
Hello, I built a computer some time ago, but started to really use it just recently. Despite having chosen parts that should be of good quality the system is not stable.

I am experiencing very strange and sudden resets of the whole computer. It is not regular at all and can be weeks between, but when it happens the computer just turns black and then starts to boot like normal. There is no Blue Screen in Windows, it is like the power just is lost. The Windows Eventlogs reports nothing connected to the problem.

The hardware setup is this:

CPU: Intel Core i7 860 (hyperthreading enabled in BIOS)
Motherboard: Asus P7P55D-E LX
Memory: 2x4 GB Corsair XMS3 DDR3 1333MHz
GPU: XFX Radeon HD5770 XT 1GB GDDR5
Power supply unit: Cooler Master Silent PRO ATX12V 2.3 600W
Harddisks: 3 x Samsung Spinpoint F3 1TB (no raid)
Case: Antec P183
Case fans: Akasa Apache Black 120mm (replaced the built in)
CPU fan: Arctic Cooling Freezer 7 PRO REV.2

The operating system is Windows 7 Enterprise 64 bit. There is no overclocking done at the CPU. I use the MSI Afterburner to lower the GPU fan speed somewhat to decrease noise.

Things that I have tried to verify:

I have run the Windows offline Memtest, it reported no errors. Also several times loaded programs (VMs) that filled nearly the whole memory and had them running for days.

I have tested the CPU with Prime95 running 8-10 hours at maximum stress with no problems.

I have verified the CPU temperatures with Cpuid Hardware Monitor and they are on average at 40 C.

I have run both online CHKDSK tests of the disks and also a very extensive offline, with no reported errors. The test included checking each sector for errors. The disks has also been stressed by running performance tools like IOmeter in Windows.

The resets has happened just a few times, and the computer could be one for a week without any problems, so it is quite infrequent. I can no see no pattern at all.
 
Solution
Not sure about windows memory tester,
but memtest the program only checks the first 64kb of a stick for errors, not the whole stick, I had a system that had random bsod's and memtest said it was fine,passed several series of tests ok, tried booting on one stick at a time and the rogue stick wouldn't boot at all, replaced it and systems fine.

a short could be fiddly to find or intermittent, are all the mobo standoffs screwed in as far as possible? is the mobo screwed firmly to them in all the right places, you may have put a standoff in the case that has no corresponding hole on the mobo, and that can short out, count the holes in your mobo, and make sure theres the same number of standoffs.

connectors 'should' be easy to connect...
Try booting your system with one stick of ram at a time, that will highlight any ram errors better than memtest,
check for shorts on the mobo,including the standoffs.
everything connected up properly? no loose cables or plugs?
psu definately ok?
might be worth running through section three on this link to doublecheck your build

http://www.tomshardware.co.uk/forum/288241-13-read-posting-buyer-guides-troubleshooting
post back if any progress/problems
Moto
 

ricno

Distinguished
Apr 5, 2010
582
0
19,010
Hello, and thanks a lot for your answer.



So the Windows memtest (or any other 3rd party memtester) is not really reliable? I am thinking that if it was a memory error, should it not crash the operating system? That is, leave a BSOD instead of this instant reboot?

I have used both normal programs and Prime95 to fill the RAM without getting the trouble, but that might not mean the memory is ok.

The slight problem also with booting with one RAM stick is that the restarts is very infrequent. I have had the computer running for over a week (and thought the problem was gone) when I got a restart last night. So it is a bit hard to actually test unfortunaly.



Would a short on the motherboard be instantly visible or could it somehow "rest" and then cause the problem? Is there any physical signs to look for when checking the motherboard?



I did a recheck earlier and made sure all cables felt like they were securly attached. Could there be any logical mistakes with the cabling that should cause this? Or would that be visible directly, as in the computer not booting?



I have actually no idea! It is brand new, but apart from that I do not know how to test or check?

Since it is the power that seems to be turned off for a brief second before the restart I do suspect some kind of electrical error.

Should the 600W be enough for my system? Could it be underpowered?
 
Not sure about windows memory tester,
but memtest the program only checks the first 64kb of a stick for errors, not the whole stick, I had a system that had random bsod's and memtest said it was fine,passed several series of tests ok, tried booting on one stick at a time and the rogue stick wouldn't boot at all, replaced it and systems fine.

a short could be fiddly to find or intermittent, are all the mobo standoffs screwed in as far as possible? is the mobo screwed firmly to them in all the right places, you may have put a standoff in the case that has no corresponding hole on the mobo, and that can short out, count the holes in your mobo, and make sure theres the same number of standoffs.

connectors 'should' be easy to connect, most are form-fit so they can only physically go in one way, its possible wires are wrongly sat in the connectors but I think that would've been apparent earlier. (more potentially destructive)

Try building your system into this, http://www.thermaltake.outervision.com/ But I'm pretty sure that 600w is enough for your build, I'm more inclined to think its a short or memory issue at this point.
Let us know how you get on,
Moto
 
Solution

ricno

Distinguished
Apr 5, 2010
582
0
19,010


When I did run the Win memtester earlier I got the impression it tested the whole memory, since I just thought that should be obvious if it not did it.

I have downloaded now the Memtest86+ tool, I shall run it today after work and see what it says.



Since it was some time ago I actually put the computer together I can not say for sure how it is done. It should be a standoff for each hole in the motherboard and no standoffs on the case that does not match to a motherboard hole, right?

Can there be anything wrong with the types of screws used to attach the motherboard to the standoffs/case? As in the width of the "hat" of the screw?

How hard shall the screws connect to the motherboard surface?

(Here is by the way some pictures of this build from another thread some time ago.)



That is good to know, then there should not reasonable be any wrongly attached cables/wires. I shall check again that they are firmly connected.



Thanks for the link. I entered all the information I had into the page and it reported that 368W should be enough. I guess the page only does some kind of good estimation, but it is still a long way from 600W!



I hope be able to better test the memory tonight with the other Memtest86+ tool and also check the standoffs. You do not suspect the PSU could be malfunctioning?
 

ricno

Distinguished
Apr 5, 2010
582
0
19,010


When I bought the PSU I did some research and asked around here at the forum and got the impression that this particular Cooler Master model should be of high quality. It could be malfunctioning never the less of course.

The model is "COOLER MASTER SILENT PRO ATX12V 2.3 600W".
 
Yup, one standoff per hole in mobo, no others should be in the case, and unless you have some deformed screws the flat part of the head shouldn't overlap the contact ring on the mobo,tight fit but if you feel resistance at any point, stop :) crushing your mobo cant be good for it I reckon
and i agree with several posters on your link, great job on the cable management, whats that molex? doing under the pcie slot though? might be worth taping that off if its not used, it could be shorting you out,
As mentioned the Cm psu's aren't top quality, but reasonable, and yes its possible you got a bad unit, but without a known good psu to swap it out with, theres not a whole lot a home tester can do with a suspect psu :(
Moto
 

ricno

Distinguished
Apr 5, 2010
582
0
19,010
Thanks again for your reply.



Which cable are you referring to? On which picture do you see it?
 

ricno

Distinguished
Apr 5, 2010
582
0
19,010


Here is a close up of the cables. Which ones do you think looks strange?

iw7yfl.jpg


 

ricno

Distinguished
Apr 5, 2010
582
0
19,010


It is a cable I found no place to connect when I built the computer. It is labeled "HD Audio". I just stuck it away under the bottom slot exit. Do you think that it could cause any problems?

2573a5u.jpg


 
Aah, should be fine, (I shoulda guessed what that was though lol) its an alternative audio connector over your ac-97 one, no problem,
Looking more and more like the psu is the culprit now tbh, do you have an old one you know works well? swap it over to check if you can, if not, see about returning that (New) one to supplier for exchange.
Moto
 

ricno

Distinguished
Apr 5, 2010
582
0
19,010
I ran the Windows Memory Tester again last night and found that I could run an extended test, that lasted for several hours. I really belive it tested the whole memory with several different methods, and it reported no errors.

I shall try the Memtest86+ later just to doublecheck.
 

ricno

Distinguished
Apr 5, 2010
582
0
19,010
Some updates:

I have verified that all standoffs (six) has a corresponding screw attached on the motherboard. I have also tighten them all up somewhat.

Last night I ran Memtest86+ for two hours and it did lots of different tests and found no errors. So neither Microsoft Windows Tester or Memtest86+ found any memory problems, so I guess the RAM could be ruled out?

Is there any possibility that the motherboard could be causing such resets? The board is a Asus P7P55D-E LX and the BIOS revision was "0401" (from 2010), but I did two updates last night, first to "1003" and then to "1208". Both these has a very limited relase notes, at least what I found, but it said that they were fixes for "system stability".
 

ricno

Distinguished
Apr 5, 2010
582
0
19,010
I have now have the computer running day and night in three weeks without any problems!

Since the resets were quite unregular and I could never produce the error I am not really, but almost, certain the problems are now gone.

However, what do you think caused this? The two things I did was to tighten up the screws on the motherboard somewhat, but none were really lose before, and I did also this BIOS update.

I do suspect the motherboard for being responsible, and the BIOS updates vaguely included fixes for "system stability", however I have found no other having these kinds of resets with this ASUS board.
 

ricno

Distinguished
Apr 5, 2010
582
0
19,010


Thanks! I will still have it to run for some time before being totally sure, but you think it could have the screws which caused this? The most strange thing was that the resets were so infrequent. When a problem happens each time you do a specific action it is also easy to try something and watch the result. :)

It would be very interesting to really know.
 
One of the standard checks in the troubleshooting link I posted is making sure everythings correctly and securely in place, plugs,screws etc, :p
loose Mobo screws can allow the Mobo to flex which could have been the problem, plus they help ground it, if you were losing ground connection the psu may have shut down for safety sake,
I thought you'd gone through the guide though so never thought to mention it,**edit, I did ask earlier on in the thread lol :p**
Moto
 

AndyC682

Distinguished
Apr 27, 2010
52
0
18,530
I had this same problem (random restarts with no BSOD's) and ended up taking it to a repair shop cos i just couldnt work it out.

I was using the same sticks as you... it ended up being the timings were out on the sticks. Trying swapping your RAM out with another kit, thats what i ended up doing and i havent had a random restart since.

 

ricno

Distinguished
Apr 5, 2010
582
0
19,010


I had already checked all other wires and connectors before starting the thread and after you suggested checking the motherboard screws I checked them out too. This was about three weeks ago. I would say the screws were reasonable firmly connected, but I remember being unsure how hard to connect them when building the PC, so it could very well been too soft.

As for losing ground as you mention, what would that mean? The computer is connected to non grounded power outlet and I have been wondering if that could be a part of the problem?




Interesting. Did you try some memory testing tools? I have both used the memory quite a lot for running many large virtual machines for a long time and also used both the Windows memory tester and the Memtest 86+ which did very extensive testing for many hours without any problems detected. However, could a timing error still exist? (I have no idea really how that could be detected.)
 

AndyC682

Distinguished
Apr 27, 2010
52
0
18,530
I had it alot more severe than you were experiencing. My rig would restart after 15-20 mins of being turned on, so i wasnt able to test myself. I took it to a repair center and they were the 1's that found the problem.

After swapping the RAM out, it hasnt happened again.
 

ricno

Distinguished
Apr 5, 2010
582
0
19,010


If your resets were so frequent I would also suspect the RAM, however for me my problems were quite unregular and I am still not sure if the motherboard screws or an error in the BIOS was responsible.

Since I have not yet had the reset again I am slowly thinking it is gone.